Linguistic universals and the order of verbs in Germanic

关于语言共性的深度和起源的争论是当前语言学的一个核心部分。他们的兴趣来自于这样一个事实，即他们对思维如何工作的问题，所有人类的多样性或相似性，以及先天和后天在其中扮演的角色。我将通过对一组特定的普遍主义建议提供进一步的经验支持来促进这些辩论。这将通过研究日耳曼语言和方言中可能的和不可能的语序模式来实现。最近的工作激发了强烈的普遍主义者对语序的主张。五（2005年：根据格林伯格的《通用20》（Universal 20）及其《语言学研究》（Linguistic Inquiry），所有被证实的中性名词（N）、形容词（A）、数词（N）和指示词（Dem）都是从一个共同的基本结构类型衍生出来的简单排列。在英语中，和许多其他语言一样，顺序是“这五个公共场所”，即，在一种较为罕见的语言中，例如班图语基库尤，指示、数词和形容词的相对顺序与英语相同，而名词排在第一位：N-Dem-Num-A。与此相反的顺序在任何语言中都没有得到证实。在Dem、Nam、A和N的24个逻辑上可能的顺序中，Cinque（2005）报告了14个被证明的顺序和10个未被证明的顺序，并提出了一个简单的、形式化的理论，该理论生成了所有且只有14个被证明的顺序。目前的项目侧重于日耳曼语言和方言。在这些语言和方言中，动词和动词相关成分的顺序是众所周知的变化。在句子“John will want to call up Sue”中，动词“will”、“want”、“call”和助词“up”的顺序是will-want-call-up，在德语中则正好相反：up-call-want-will。英语和德语之间的这种对比让人想起上面提到的英语和冈贝语之间的对比。在荷兰语中，这个顺序是“up-will-want-call”，其中“will”、“want”和“call”的顺序类似于英语，但小品词在前面。这让人想起了名词短语中的基库尤顺序，这是英语式的，除了名词在开头的位置。荷兰语秩序的镜像，“call-want-will-up”，在任何语言或方言中都找不到。这再次让人想起，在关于名词短语的跨语言记录中，没有出现反向的基库尤语序。日耳曼语中名词短语和动词之间的排序可能性之间的相似性是惊人的和有启发性的。我将首先研究，在日耳曼语言和方言中发现的动词之间的可变顺序在多大程度上可以用上面描述的名词短语的顺序来正式表征。本研究围绕着一个实证问题展开，即在四个动词和动词相关成分组合在一起的构式中，是否所有且只有14个预期的语序出现在日耳曼语的某个变体中。如果这一假设能够成立，那么日耳曼语中动词的语序就不能再被视为辛克著作中的一个关键建议的反例：在中性语序的生成过程中，每当一个项目（如基库尤中的名词或荷兰语中的助词）移动时，它就向前移动。这样的结果将大大加强普遍主义者的立场。第二组问题涉及关于动词要素顺序的更微妙的概括。例如，虽然日耳曼语言和方言之间的词序有很大的差异，但没有一种语言或方言可以在分词之后跟随层次上更突出的元素，而不定式必须在这种层次上更突出的元素之前。这种从文献中归纳出来的结论将被经验性地检验，并在结构方面寻求解释。

Debates over the depth and provenance of linguistic universals are a central part of current linguistics. Their interest derives from the fact that they bear on the issue of how mind works, how diverse or similar all humans are, and what roles nature and nurture play in this. I will contribute to these debates by giving further empirical support to a particular set of universalist proposals. This will be achieved by studying possible and impossible word order patterns within the Germanic languages and dialects.Recent work motivates strong universalist claims about word order. Cinque (2005: Deriving Greenberg's Universal 20 and its Exceptions, Linguistic Inquiry) has argued that all attested neutral orders of nouns (N), adjectives (A), numerals (Num), and demonstratives (Dem) are simple permutations derived from a common underlying structural type. In English, as in many other languages, the order is 'these five public employees', i.e., Dem-Num-A-N, while in the Kwa language Gungbe, as again in many other languages, it is the exact opposite: N-A-Num-Dem. In a rarer type of language, for example the Bantu language Kikuyu, the relative order of demonstrative, numeral and adjective is the same as in English, while the noun comes first: N-Dem-Num-A. The exact opposite of this order is not attested in any language. Of the 24 logically possible orders of Dem, Num, A, and N, Cinque (2005) reports 14 as attested and 10 as unattested and suggests a simple, formal theory that generates all and only the 14 attested orders. The current project focuses on Germanic languages and dialects. In these languages and dialects, the order of verbs and verb-related elements is notoriously variable. The order of the verbs 'will', 'want', 'call' and the particle 'up' in the sentence 'John will want to call up Sue' is will-want-call-up, it would be the exact opposite in German: up-call-want-will. This contrast between English and German is reminiscent of that between English and Gungbe mentioned above. In Dutch the order comes out as 'up-will-want-call', where 'will', 'want', and 'call' come in an English-like order, but the particle precedes. This is reminiscent of the Kikuyu order in the noun phrase, which was English-like, except with the noun in initial position. The mirror image of the Dutch order, 'call-want-will-up', is never found in any language or dialect. This is again reminiscent of the absence of the reverse Kikuyu order from the cross-linguistic record regarding the noun phrase. The similarity between ordering possibilities in the noun phrase and among verbs in Germanic are striking and suggestive. I will investigate first, to what extent the variable orders found amongst verbs in the Germanic languages and dialects can be characterized formally in the same terms as the orders described above for the noun phrase. The research revolves around the empirical question whether in constructions that cluster together four verb and verb-related elements all and only the 14 expected orders show up in some Germanic variant or other. If this hypothesis can be upheld, the orders of verbs in Germanic can no longer be seen as a counterexample to one of the crucial proposals contained in Cinque's work: In the generation of neutral word orders, whenever an item (such as the noun in Kikuyu or the particle in Dutch) moves, it moves forward. Such a result would strengthen the universalist position considerably. The second set of questions concerns more subtle generalizations regarding the order of verbal elements. For example, while there are large word-order differences between the Germanic languages and dialects, there is no language or dialect where a participle could follow a hierarchically more prominent element while infinitives have to precede such hierarchically more prominent elements. Such generalizations from the literature willbe tested empirically and an explanation sought in structural terms.